Reproduce “SAFARI: Semantic- And Frontier-driven Autonomous Risk Inspection” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Ugur Kafkasyali, Mihnea Toader, Julia Richter, Turcan Tuna, Cesar Cadena, Marco Hutter

SAFARI uses a persistent risk map to guide robot reconnaissance for mission-relevant objects. It combines depth-supported detections with visibility-based non-detections conditioned on assumed object size, using uncertainty to direct observation and risk-aware motion planning. Across 11 field missions, average coverage was 98.7% of accessible area, not hazards found; in one road test, a larger-size setting reduced detections from 8/8 to 4/8.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.